The Scandal Isn’t Over

The federal government’s insurance website is working. We should still be furious about its failures. is working now, but it’s still a rip-off.

Screenshot of This time, it’s not a disaster! While there have been some hiccups in the Obamacare open enrollment period that began Nov. 15, we’ve seen nothing like the nonfunctioning calamity that accompanied the initial rollout on Oct. 1 of last year. That launch revealed severe problems with government contracting, the management of the United States Department of Health and Human Services and the Centers for Medicare and Medicaid Services, and accountability in general. Dozens of contractors, with about a half-dozen predominating, were paid about $800 million by the government to produce what was essentially half-finished garbage. Stories of people successfully enrolling in Obamacare were next to nonexistent in the first month of the rollout. With contractors shifting blame in front of Congress, Republicans looking to score points without caring to salvage the situation, Democrats playing weak defense with HHS, and CMS reshuffling PowerPoints, the fiasco was only rectified once the Obama administration recruited Jeffrey Zients as a crisis manager. He recruited experts from Google and elsewhere and more or less pulled things together by the end of November.

A year later, the crisis is long over. Now’s the time for what developers call a postmortem: figuring out what went wrong, how it went wrong, and how to prevent it from happening again. was the first high-profile Web tech initiative from the federal government, and it was a disaster. With a smooth rollout success rate of 0 percent, has the government learned from its mistakes, or will the next ambitious tech project also sputter? Without access to government internals, the analysis is inconclusive, but there are a few telling clues, not all of them good.

Around the time Zients took over, Anderson Cooper asked me on AC360 Later to rate the difficulty of fixing the problems. While I gave the purely technical side just 4 out of 10, the bureaucratic and managerial challenges rated a full 10 out of 10. My reason was that the full functioning of required coordination between literally hundreds of entities, from federal agencies to state governments to insurance companies. With the federal work initially parceled out to at least a half-dozen different contracting firms, no one in the government or in contracting had been responsible for making sure the entire thing worked end-to-end before release. And clearly anyone who bothered to make clear that a fiasco awaited (as any competent programmer would have known) was ignored.

The blame has mostly been cast at former HHS secretary Kathleen Sebelius, who resigned over the rollout failure, as well as Canadian contractor CGI Federal, which was responsible for the (non)integration of the website and was axed by the administration in favor of another key contractor, QSSI, owned by health insurer UnitedHealth. Former QSSI vice president Andrew Slavitt has, in fact, joined the administration as CMS’ principal deputy administrator, second in command under administrator Marilyn Tavenner. I was harshly critical of Slavitt’s performance in front of Congress last year, saying that he only looked good in comparison with the shameless buck-passing of CGI’s Cheryl Campbell, who had repeatedly fingered CMS as the culprit at the congressional hearings. When asked about the security of health care data, Slavitt gaffed, “Our systems don’t hold data. They just transport data through it.” On July 11, Slavitt was granted an ethics waiver by the White House so he could transfer directly from QSSI to CMS without having to wait a year, since the job would necessarily entail dealing with UnitedHealth. The waiver trumpets his skills without mentioning that he had anything to do with prior to September 2013. “CMS has informed me that they are unaware of any other candidate with this unique blend of skills and knowledge who is available and willing to take on this challenging and critical position,” writes Edgar Swindell, associate general counsel for ethics of HHS. If that’s true, we may well be doomed. QSSI, whose touted “data hub” also failed to perform as expected on Oct. 1, 2013, took over as primary contractor during the crisis period, albeit under direction from Zients’ team. UnitedHealth got its man on the inside, with his company’s role in the fiasco and his poor congressional performance posing no obstacle. It’s possible that Slavitt is more talented than his résumé would suggest, but even if you give him the benefit of the doubt there, the conflict of interest and the sheer optics make him a poor choice.

Meanwhile, CGI’s responsibilities were handed over in January to Accenture, known as a competent but expensive tech integration contractor. (The Washington Post has chronicled Accenture’s history of ethical lapses and failed projects.) Awarded a $121 million contract for what is essentially a redo of, Accenture seems not to have fallen down on the job, but that’s $121 million that should never have needed to be spent. If CGI refunded any of its take, I never heard about it. Maybe CGI gave the administration a break on the multibillion-dollar unlimited ceiling umbrella contract vehicle known as One Acquisition Solution for Integrated Services, or OASIS, awarded to CGI in July by the administration. CGI Federal President James Peake said, “It is a true reflection of our successful past performance delivering high-quality professional services to our clients,” apparently while keeping a straight face. OASIS is what’s known as an “indefinite delivery indefinite quantity” contract, which admittedly does sound perfect for CGI. Covering everything from defense to finance to engineering, OASIS appears to be an incomprehensible, consultant-driven “innovation” wrapped in empty jargon, such as this list of OASIS “features and benefits” provided by GSA:      

  • Government-wide use
  • Access to best in class solution providers
  • On-ramp/off-ramp procedures to ensure a flexible, vibrant vendor pool
  • Integrated support for key government initiatives

That’s not all CGI got away with. Vermont, whose state-run exchange was also built by CGI and was also a nonfunctional disaster, only disposed of CGI this August, an inexplicable delay that left CGI with a fat $67 million check for building terrible software that was delivered with continual delays and overruns. (Five million dollars in penalties were assessed, however.) Vermonters should feel outraged over the sheer waste and indifference to accountability that their state government has shown. I can comfort them with the irony that their new tech contractor is Optum, owner of QSSI and a subsidiary of UnitedHealth, former employer of CMS Deputy Administrator Andrew Slavitt. It’s all in the family.

Even if the moderate technical challenge of had been met, the cost still would have been exorbitant. The health care website cost about $800 million to build, double the original estimate of $400 million. CGI got $250 million of that, while QSSI got $165 million. (For comparison, the entire development cost for the revolutionary first iteration of the iPhone was just $150 million.) That is vastly beyond the market cap of most startups, much less their actual R&D expenses. Despite paying well more than 10 times what it should have, the government got a broken website, but even if it had worked, it would have been a tremendous rip-off. There appears to be little interest by anyone in looking into the fat contracts handed out to contractors to do work that ranges from moderately able to totally incompetent, as well as the contracting rules that prohibit anyone but a small number of companies from even being able to get those juicy contracts in the first place. If greater accountability, transparency, and efficiency were brought to the entire process of federal tech contracting (and, perhaps, contracting in general), the government savings would be on the order of billions. Despite the Obama administration announcing a “Campaign to Cut Waste” back in 2011 with “Reform Government Contracting” as a key goal, there is little evidence that reform has occurred, even after the catastrophe. All we’ve seen is a game of musical chairs.